Cost-Aware Distributed Online Learning with Strict Rejection Behavior against Adversarial Agents

Runqi Chai; Senchun Chai; Xudong Zhao; Yuanqing Xia; Yuhan Suo

arxiv: 2412.01524 · v6 · pith:L2T4ISZZnew · submitted 2024-12-02 · 💻 cs.MA · cs.SI· math.OC

Cost-Aware Distributed Online Learning with Strict Rejection Behavior against Adversarial Agents

Yuhan Suo , Runqi Chai , Senchun Chai , Xudong Zhao , Yuanqing Xia This is my paper

Pith reviewed 2026-05-25 08:22 UTC · model grok-4.3

classification 💻 cs.MA cs.SImath.OC

keywords distributed online learningadversarial agentsstrict rejectionevolution desynchronizationtwo-time-scale architecturemulti-agent systemsIoT monitoring

0 comments

The pith

Strict rejection of adversarial agents induces desynchronization that a two-time-scale regulator can attenuate in distributed learning.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a cost-aware distributed online learning framework that uses strict rejection to block adversarial agents from influencing normal ones. It identifies that this rejection can cause neighboring normal agents to evolve at different rates temporarily, leading to desynchronization across the network. To fix this, the authors add a two-time-scale architecture with an outer layer that adjusts long-term rates and an inner layer for learning. This setup is important because existing methods overlook the inefficiency from repeated corrections in systems like IoT multi-agent networks facing ongoing threats.

Core claim

The central claim is that the proposed two-time-scale adaptive evolution regulation architecture attenuates the propagation of strict-rejection-induced evolution desynchronization while the outer layer maintains a dynamic tracking property, enabling robust and low-cost distributed online learning under persistent malicious interference.

What carries the argument

The two-time-scale adaptive evolution regulation architecture that dynamically adjusts the long-term evolution-rate schedule in its outer layer while the inner layer preserves robust online learning.

Load-bearing premise

Strict rejection of adversarial agents induces heterogeneous transient evolution among neighboring normal agents, creating a desynchronization problem that requires the two-time-scale architecture to mitigate.

What would settle it

A simulation or experiment where the two-time-scale regulation is applied but the desynchronization across the network does not decrease compared to the case without regulation.

Figures

Figures reproduced from arXiv: 2412.01524 by Runqi Chai, Senchun Chai, Xudong Zhao, Yuanqing Xia, Yuhan Suo.

**Figure 2.** Figure 2: Opinion evolution input cost with and without the pro [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Figure 3: Opinion evolution input cost with different [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

**Figure 4.** Figure 4: The weight matrix ωij (k) changes for normal agents towards neighbors. [7] C.-X. Shi and G.-H. Yang, “Secure bearing-based target localization for multi-agent networks against malicious agents,” IEEE Transactions on Automation Science and Engineering, 2023. [8] H. Guo, Z.-H. Pang, J. Sun, and J. Li, “An output-coding-based detection scheme against replay attacks in cyber-physical systems,” IEEE Transaction… view at source ↗

read the original abstract

Distributed online learning in Internet of Things(IoT)-enabled multi-agent systems(MASs) is highly vulnerable to persistent adversarial interactions, particularly when malicious agents cannot be fully isolated during the transient learning stage. Existing resilient learning methods mainly focus on convergence preservation or malicious suppression, while the resulting evolution inefficiency caused by repeated corrective adaptation remains largely unexplored. To address this issue, this paper develops a cost-aware distributed online learning framework with a strict rejection behavior against adversarial agents. The proposed mechanism suppresses harmful assimilation of suspicious neighboring information and reveals a previously overlooked side effect, that is, the strict rejection may induce heterogeneous transient evolution among neighboring normal agents, leading to evolution desynchronization across the network. To mitigate this effect, a two-time-scale adaptive evolution regulation architecture is further developed, in which the outer layer dynamically adjusts the long-term evolution-rate schedule while the inner layer preserves robust online learning. Theoretical analysis establishes the dynamic tracking property of the outer-layer update and proves that the proposed regulation mechanism attenuates the propagation of strict-rejection-induced evolution desynchronization. Numerical simulations and a satellite-assisted IoT monitoring scenario demonstrate that the proposed method achieves robust and low-cost distributed online learning under persistent malicious interference.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper flags a desync side-effect from strict rejection in adversarial MAS learning and counters it with a two-time-scale regulator; the fix is technically coherent but narrow in payoff.

read the letter

The main contribution here is spotting that strict rejection of adversarial neighbors can leave normal agents evolving at different rates, then adding an outer slow layer to adjust long-term schedules while the inner layer keeps the online updates. That framing and the two-time-scale split are new relative to prior resilient methods that only chased convergence or isolation. The Lyapunov-plus-singular-perturbation argument for the outer-layer tracking property holds together under the stated connectivity and separation assumptions, and the satellite IoT simulation shows the expected attenuation of the desync effect without obvious overfitting. Credit for shipping both the analysis and a concrete scenario instead of stopping at abstract claims. The soft spots are modest but real: the connectivity and attack models are standard for this subfield, so the result stays inside the MAS resilient-learning lane rather than generalizing; the cost-awareness piece feels more like an add-on than a driver of the new architecture; and the numerical evidence is illustrative rather than a stress test across varied graph sizes or attack intensities. A reader already working on distributed online learning with persistent adversaries will find the side-effect diagnosis and the regulator useful; someone outside that niche will not. The work is internally consistent and shows clear engagement with the literature, so it deserves a serious referee rather than a desk reject.

Referee Report

0 major / 3 minor

Summary. The paper proposes a cost-aware distributed online learning framework for IoT-enabled multi-agent systems that incorporates strict rejection of adversarial agents to suppress harmful assimilation. It identifies an overlooked side effect whereby strict rejection induces heterogeneous transient evolution and desynchronization among normal agents, and introduces a two-time-scale adaptive regulation architecture (outer layer for long-term evolution-rate scheduling, inner layer for robust learning) to mitigate it. Theoretical analysis establishes the outer-layer dynamic tracking property and proves attenuation of strict-rejection-induced desynchronization propagation, with supporting numerical simulations and a satellite-assisted IoT monitoring case study.

Significance. If the central claims hold, the work makes a useful contribution by jointly addressing security against persistent adversaries and the resulting efficiency costs in distributed online learning, an area where prior resilient methods have focused mainly on convergence preservation. The explicit treatment of the desynchronization side effect and the two-time-scale separation (analyzed via standard Lyapunov and singular-perturbation techniques) provide a concrete mechanism that could improve practical performance in adversarial MAS settings such as IoT monitoring.

minor comments (3)

[§3.2] §3.2, the definition of the strict-rejection threshold and the cost function: the notation for the indicator function and the weighting parameter could be clarified to avoid ambiguity with the inner-layer update rule.
[Figure 4] Figure 4 and the accompanying simulation description: the caption does not explicitly state the number of Monte Carlo runs or the exact adversarial injection schedule used to generate the plotted trajectories.
[References] The reference list omits several recent works on two-time-scale MAS coordination that would help situate the regulation architecture relative to existing singular-perturbation results.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the constructive and positive review, which accurately captures the paper's contributions on cost-aware distributed online learning with strict adversarial rejection and the two-time-scale regulation to mitigate desynchronization. The recommendation for minor revision is noted, and we will incorporate improvements for clarity and presentation in the revised version.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper's derivation relies on standard Lyapunov stability analysis and singular perturbation techniques applied to a two-time-scale architecture under explicit MAS connectivity and model assumptions. The dynamic tracking property and desynchronization attenuation are established as consequences of these techniques rather than by redefinition or fitting. No self-citation chains, ansatzes smuggled via prior work, or predictions that reduce to fitted inputs appear in the load-bearing steps; the side-effect premise is motivated directly from the strict-rejection mechanism without circular reduction to the target result.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Assessment limited to abstract only; no explicit free parameters, axioms, or invented entities are identifiable. The work likely rests on standard assumptions from distributed optimization and online learning literature, but these cannot be audited without the full text.

pith-pipeline@v0.9.0 · 5753 in / 1135 out tokens · 38044 ms · 2026-05-25T08:22:09.526434+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Theorem 3.1: ΔJ_i(k) ≤ λ_max(G_i(k)) (1−γ)^k δ_i²(k) with G_i = S^T (K^T R K) S; cost Ji(u) = ∑ (1−γ)^{-k} u^T R u minimized by PDARE solution P_i

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Robust Set Partitioning Strategy for Malicious Information Detection in Large-Scale Internet of Things
cs.DC 2025-02 unverdicted novelty 5.0

A new set partitioning strategy using Grassmann distance and a gain mutual influence metric enables distributed attack detection in large IoT networks with at most 1.648% performance gap and O(1/m) computation reduction.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages · cited by 1 Pith paper

[1]

Marp: A cooperative multi-agent drl system for connected a utonomous vehicle platooning,

S. Dai, S. Li, H. Tang, X. Ning, F. Fang, Y . Fu, Q. Wang, and L . Cheng, “Marp: A cooperative multi-agent drl system for connected a utonomous vehicle platooning,” IEEE Internet of Things Journal , 2024

work page 2024
[2]

Product diffusion in dynami c online social networks: A multi-agent simulation based on gravity theory,

X. Wei, H. Gong, and L. Song, “Product diffusion in dynami c online social networks: A multi-agent simulation based on gravity theory,” Expert Systems with Applications , vol. 213, p. 119008, 2023

work page 2023
[3]

Swif t: A distributed one-stage planner for efﬁcient multi-quadrot or trajectory op- timization,

H. Wang, S. Zhang, Y . Sun, Z. Wang, J. Sun, and B. Zhu, “Swif t: A distributed one-stage planner for efﬁcient multi-quadrot or trajectory op- timization,” IEEE Transactions on Automation Science and Engineering , vol. 22, pp. 20 951–20 965, 2025

work page 2025
[4]

Consensus in multi-age nt systems: a review,

A. Amirkhani and A. H. Barshooi, “Consensus in multi-age nt systems: a review,” Artiﬁcial Intelligence Review , vol. 55, no. 5, pp. 3897–3935, 2022

work page 2022
[5]

Fundamental perfor mance limitations for average consensus in open multi-agent syst ems,

C. M. de Galland and J. M. Hendrickx, “Fundamental perfor mance limitations for average consensus in open multi-agent syst ems,” IEEE Transactions on Automatic Control , vol. 68, no. 2, pp. 646–659, 2022

work page 2022
[6]

Con sensus- based distributed connectivity control in multi-agent sys tems,

K. Griparic, M. Polic, M. Krizmancic, and S. Bogdan, “Con sensus- based distributed connectivity control in multi-agent sys tems,” IEEE transactions on network science and engineering , vol. 9, no. 3, pp. 1264–1281, 2022. /s48 /s53 /s49 /s48 /s49 /s53 /s50 /s48 /s50 /s53 /s51 /s48 /s51 /s53 /s52 /s48 /s52 /s53 /s53 /s48 /s49 /s50 /s51 /s52 /s53 /s54 /s78/...

work page 2022
[7]

Secure bearing-based target l ocalization for multi-agent networks against malicious agents,

C.-X. Shi and G.-H. Y ang, “Secure bearing-based target l ocalization for multi-agent networks against malicious agents,” IEEE Transactions on Automation Science and Engineering , 2023

work page 2023
[8]

An output-coding-b ased detection scheme against replay attacks in cyber-physical systems,

H. Guo, Z.-H. Pang, J. Sun, and J. Li, “An output-coding-b ased detection scheme against replay attacks in cyber-physical systems,” IEEE Transactions on Circuits and Systems II: Express Briefs , vol. 68, no. 10, pp. 3306–3310, 2021

work page 2021
[9]

Resilient consensus control for heteroge- neous multiagent systems via multiround attack detection a nd isolation algorithm,

W. Y ue, Y . Y ang, and W. Sun, “Resilient consensus control for heteroge- neous multiagent systems via multiround attack detection a nd isolation algorithm,” IEEE Transactions on Industrial Informatics , 2023

work page 2023
[10]

Exploiting trust for resilient hypothesis testing with ma licious robots,

M. Cavorsi, O. E. Akg¨ un, M. Y emini, A. J. Goldsmith, and S. Gil, “Exploiting trust for resilient hypothesis testing with ma licious robots,” in 2023 IEEE International Conference on Robotics and Automat ion (ICRA). IEEE, 2023, pp. 7663–7669

work page 2023
[11]

Fast distributed optimization over directed graphs under malic ious attacks using trust,

A. K. Dayı, O. E. Akg¨ un, S. Gil, M. Y emini, and A. Nedi´ c, “Fast distributed optimization over directed graphs under malic ious attacks using trust,” arXiv preprint arXiv:2407.06541 , 2024

work page arXiv 2024
[12]

Resilient freq uency estimation for renewable power generation against phasor m easurement unit and communication link failures,

Z. Hu, R. Su, K. Zhang, R. Wang, and R. Ma, “Resilient freq uency estimation for renewable power generation against phasor m easurement unit and communication link failures,” IEEE Transactions on Circuits and Systems II: Express Briefs , vol. 72, no. 1, pp. 233–237, 2025

work page 2025
[13]

Security analysis and defense strategy of distributed ﬁltering unde r false data injection attacks,

J. Zhou, W. Y ang, H. Zhang, W. X. Zheng, Y . Xu, and Y . Tang, “Security analysis and defense strategy of distributed ﬁltering unde r false data injection attacks,” Automatica, vol. 138, p. 110151, 2022

work page 2022
[14]

Distribute d estimation with cross-veriﬁcation under false data-injection attack s,

Y . Hua, F. Wan, H. Gan, Y . Zhang, and X. Qing, “Distribute d estimation with cross-veriﬁcation under false data-injection attack s,” IEEE Trans- actions on Cybernetics , vol. 53, no. 9, pp. 5840–5853, 2023

work page 2023
[15]

A reputation awareness randomization consensus mechanism i n blockchain systems,

J. Zhang, Y . Sun, D. Guo, L. Luo, L. Li, Q. Nian, S. Zhu, and F. Y ang, “A reputation awareness randomization consensus mechanism i n blockchain systems,” IEEE Internet of Things Journal , vol. 11, no. 20, pp. 32 745– 32 758, 2024

work page 2024
[16]

Converter- based moving target defense against deception attacks in dc microgrids,

M. Liu, C. Zhao, Z. Zhang, R. Deng, P . Cheng, and J. Chen, “ Converter- based moving target defense against deception attacks in dc microgrids,” IEEE Transactions on Smart Grid , vol. 13, no. 5, pp. 3984–3996, 2022

work page 2022
[17]

The boomerang effect a synthesi s of ﬁndings and a preliminary theoretical framework,

S. Byrne and P . S. Hart, “The boomerang effect a synthesi s of ﬁndings and a preliminary theoretical framework,” Annals of the International Communication Association , vol. 33, no. 1, pp. 3–37, 2009

work page 2009
[18]

Internet, social media and online hate speech. sys tematic review,

S. A. Casta˜ no-Pulgar´ ın, N. Su´ arez-Betancur, L. M. T. V ega, and H. M. H. L ´ opez, “Internet, social media and online hate speech. sys tematic review,” Aggression and violent behavior , vol. 58, p. 101608, 2021

work page 2021
[19]

Finan cial fraud detection through the application of machine learning tech niques: a literature review,

L. Hernandez Aros, L. X. Bustamante Molano, F. Gutierre z-Portela, J. J. Moreno Hernandez, and M. S. Rodr´ ıguez Barrero, “Finan cial fraud detection through the application of machine learning tech niques: a literature review,” Humanities and Social Sciences Communications , vol. 11, no. 1, pp. 1–22, 2024

work page 2024
[20]

Epidemic spreading over m ulti-layer networks with stubborn agents,

X. Lin, Y . Shang, and Q. Jiao, “Epidemic spreading over m ulti-layer networks with stubborn agents,” IEEE Transactions on Circuits and Systems II: Express Briefs , vol. 71, no. 2, pp. 812–816, 2024. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 7

work page 2024
[21]

Cat: A consensus-adaptive trust management based on the gr oup decision making in iovs,

Y . Song, Y . Cao, C. Cheong, D. He, K.-K. Raymond Choo, and J. Wang, “Cat: A consensus-adaptive trust management based on the gr oup decision making in iovs,” IEEE Transactions on Information F orensics and Security , vol. 19, pp. 7730–7743, 2024

work page 2024
[22]

Locke, An essay concerning human understanding

J. Locke, An essay concerning human understanding . Kay & Troutman, 1847

work page
[23]

Zhou, Truncated predictor feedback for time-delay systems

B. Zhou, Truncated predictor feedback for time-delay systems. Springer, 2014

work page 2014

[1] [1]

Marp: A cooperative multi-agent drl system for connected a utonomous vehicle platooning,

S. Dai, S. Li, H. Tang, X. Ning, F. Fang, Y . Fu, Q. Wang, and L . Cheng, “Marp: A cooperative multi-agent drl system for connected a utonomous vehicle platooning,” IEEE Internet of Things Journal , 2024

work page 2024

[2] [2]

Product diffusion in dynami c online social networks: A multi-agent simulation based on gravity theory,

X. Wei, H. Gong, and L. Song, “Product diffusion in dynami c online social networks: A multi-agent simulation based on gravity theory,” Expert Systems with Applications , vol. 213, p. 119008, 2023

work page 2023

[3] [3]

Swif t: A distributed one-stage planner for efﬁcient multi-quadrot or trajectory op- timization,

H. Wang, S. Zhang, Y . Sun, Z. Wang, J. Sun, and B. Zhu, “Swif t: A distributed one-stage planner for efﬁcient multi-quadrot or trajectory op- timization,” IEEE Transactions on Automation Science and Engineering , vol. 22, pp. 20 951–20 965, 2025

work page 2025

[4] [4]

Consensus in multi-age nt systems: a review,

A. Amirkhani and A. H. Barshooi, “Consensus in multi-age nt systems: a review,” Artiﬁcial Intelligence Review , vol. 55, no. 5, pp. 3897–3935, 2022

work page 2022

[5] [5]

Fundamental perfor mance limitations for average consensus in open multi-agent syst ems,

C. M. de Galland and J. M. Hendrickx, “Fundamental perfor mance limitations for average consensus in open multi-agent syst ems,” IEEE Transactions on Automatic Control , vol. 68, no. 2, pp. 646–659, 2022

work page 2022

[6] [6]

Con sensus- based distributed connectivity control in multi-agent sys tems,

K. Griparic, M. Polic, M. Krizmancic, and S. Bogdan, “Con sensus- based distributed connectivity control in multi-agent sys tems,” IEEE transactions on network science and engineering , vol. 9, no. 3, pp. 1264–1281, 2022. /s48 /s53 /s49 /s48 /s49 /s53 /s50 /s48 /s50 /s53 /s51 /s48 /s51 /s53 /s52 /s48 /s52 /s53 /s53 /s48 /s49 /s50 /s51 /s52 /s53 /s54 /s78/...

work page 2022

[7] [7]

Secure bearing-based target l ocalization for multi-agent networks against malicious agents,

C.-X. Shi and G.-H. Y ang, “Secure bearing-based target l ocalization for multi-agent networks against malicious agents,” IEEE Transactions on Automation Science and Engineering , 2023

work page 2023

[8] [8]

An output-coding-b ased detection scheme against replay attacks in cyber-physical systems,

H. Guo, Z.-H. Pang, J. Sun, and J. Li, “An output-coding-b ased detection scheme against replay attacks in cyber-physical systems,” IEEE Transactions on Circuits and Systems II: Express Briefs , vol. 68, no. 10, pp. 3306–3310, 2021

work page 2021

[9] [9]

Resilient consensus control for heteroge- neous multiagent systems via multiround attack detection a nd isolation algorithm,

W. Y ue, Y . Y ang, and W. Sun, “Resilient consensus control for heteroge- neous multiagent systems via multiround attack detection a nd isolation algorithm,” IEEE Transactions on Industrial Informatics , 2023

work page 2023

[10] [10]

Exploiting trust for resilient hypothesis testing with ma licious robots,

M. Cavorsi, O. E. Akg¨ un, M. Y emini, A. J. Goldsmith, and S. Gil, “Exploiting trust for resilient hypothesis testing with ma licious robots,” in 2023 IEEE International Conference on Robotics and Automat ion (ICRA). IEEE, 2023, pp. 7663–7669

work page 2023

[11] [11]

Fast distributed optimization over directed graphs under malic ious attacks using trust,

A. K. Dayı, O. E. Akg¨ un, S. Gil, M. Y emini, and A. Nedi´ c, “Fast distributed optimization over directed graphs under malic ious attacks using trust,” arXiv preprint arXiv:2407.06541 , 2024

work page arXiv 2024

[12] [12]

Resilient freq uency estimation for renewable power generation against phasor m easurement unit and communication link failures,

Z. Hu, R. Su, K. Zhang, R. Wang, and R. Ma, “Resilient freq uency estimation for renewable power generation against phasor m easurement unit and communication link failures,” IEEE Transactions on Circuits and Systems II: Express Briefs , vol. 72, no. 1, pp. 233–237, 2025

work page 2025

[13] [13]

Security analysis and defense strategy of distributed ﬁltering unde r false data injection attacks,

J. Zhou, W. Y ang, H. Zhang, W. X. Zheng, Y . Xu, and Y . Tang, “Security analysis and defense strategy of distributed ﬁltering unde r false data injection attacks,” Automatica, vol. 138, p. 110151, 2022

work page 2022

[14] [14]

Distribute d estimation with cross-veriﬁcation under false data-injection attack s,

Y . Hua, F. Wan, H. Gan, Y . Zhang, and X. Qing, “Distribute d estimation with cross-veriﬁcation under false data-injection attack s,” IEEE Trans- actions on Cybernetics , vol. 53, no. 9, pp. 5840–5853, 2023

work page 2023

[15] [15]

A reputation awareness randomization consensus mechanism i n blockchain systems,

J. Zhang, Y . Sun, D. Guo, L. Luo, L. Li, Q. Nian, S. Zhu, and F. Y ang, “A reputation awareness randomization consensus mechanism i n blockchain systems,” IEEE Internet of Things Journal , vol. 11, no. 20, pp. 32 745– 32 758, 2024

work page 2024

[16] [16]

Converter- based moving target defense against deception attacks in dc microgrids,

M. Liu, C. Zhao, Z. Zhang, R. Deng, P . Cheng, and J. Chen, “ Converter- based moving target defense against deception attacks in dc microgrids,” IEEE Transactions on Smart Grid , vol. 13, no. 5, pp. 3984–3996, 2022

work page 2022

[17] [17]

The boomerang effect a synthesi s of ﬁndings and a preliminary theoretical framework,

S. Byrne and P . S. Hart, “The boomerang effect a synthesi s of ﬁndings and a preliminary theoretical framework,” Annals of the International Communication Association , vol. 33, no. 1, pp. 3–37, 2009

work page 2009

[18] [18]

Internet, social media and online hate speech. sys tematic review,

S. A. Casta˜ no-Pulgar´ ın, N. Su´ arez-Betancur, L. M. T. V ega, and H. M. H. L ´ opez, “Internet, social media and online hate speech. sys tematic review,” Aggression and violent behavior , vol. 58, p. 101608, 2021

work page 2021

[19] [19]

Finan cial fraud detection through the application of machine learning tech niques: a literature review,

L. Hernandez Aros, L. X. Bustamante Molano, F. Gutierre z-Portela, J. J. Moreno Hernandez, and M. S. Rodr´ ıguez Barrero, “Finan cial fraud detection through the application of machine learning tech niques: a literature review,” Humanities and Social Sciences Communications , vol. 11, no. 1, pp. 1–22, 2024

work page 2024

[20] [20]

Epidemic spreading over m ulti-layer networks with stubborn agents,

X. Lin, Y . Shang, and Q. Jiao, “Epidemic spreading over m ulti-layer networks with stubborn agents,” IEEE Transactions on Circuits and Systems II: Express Briefs , vol. 71, no. 2, pp. 812–816, 2024. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 7

work page 2024

[21] [21]

Cat: A consensus-adaptive trust management based on the gr oup decision making in iovs,

Y . Song, Y . Cao, C. Cheong, D. He, K.-K. Raymond Choo, and J. Wang, “Cat: A consensus-adaptive trust management based on the gr oup decision making in iovs,” IEEE Transactions on Information F orensics and Security , vol. 19, pp. 7730–7743, 2024

work page 2024

[22] [22]

Locke, An essay concerning human understanding

J. Locke, An essay concerning human understanding . Kay & Troutman, 1847

work page

[23] [23]

Zhou, Truncated predictor feedback for time-delay systems

B. Zhou, Truncated predictor feedback for time-delay systems. Springer, 2014

work page 2014